Abstract
Using a meta-analytic approach, the authors examined the predictive validity of scores on the MacAndrew Alcoholism Scale. Across 161 studies involving a total sample of 31,320 individuals, the proportion correct classification was .70. When a cut score of 24 was used across 63 studies involving a total sample of 11,113 individuals, the proportion correct classification was .72. The proportion correct classifications for the different cut scores were cumulated separately, and the functional graph suggested that the best discrimination (.76) occurred at a cut score of 27. It is argued that given the potential personality basis for these predictions, use of continuous measures rather than cut scores is preferable.
Get full access to this article
View all access options for this article.
